Registered: Sep 1999
Location: Lewisburg, PA
Posts: 56

screen, size, vs buttons

I really do use the buttons so I would hate to see them disappear. however; i think the buttons could go on the side like some of the ce machines if that would shrink the final product a bit,.

also; i think the grafitti area would be nice if it could come and go as applications need it. But, as a general rule; i think the buttons need to stay. i think the buttons are the main reason the palm is so much easier to use than ce.

how about make the next palm 160x240

(or even better 320x480) let old apps use a 4 pixel in place of one. i bet palm could find/make a 50mhz cpu compatible in the most general sense w/ the OS.

also; would anyone be that sad if a singinficant improvement in the OS cost us some compatibility isues? seriously; a 50mhz cpu could EMULATE the 15mhz dragonball if they want backwards compatibility.

Palm seems stuck. they need to keep there design principles since they are what made them succesful; but there obsolete architecture is in the way of the future it seems. I guess by making windows CE so open ended (read inefficient) they do make it more future friendly :-(

palm should write a new os, not 4.0 but a major rewrite; and throw in emulation to keep market.

------------------
Tom Alphin III - [email protected] - www.students.bucknell.edu/alphin/

Registered: Sep 1999
Location: Rio Rancho, NM, USA
Posts: 129

My only problem with a higher screen resolution is that the pixels are pretty small already. With the exception of graphics, what would you use the higher res for?

Registered: Sep 1999
Location: Urbana IL
Posts: 144

I wouldn't worry too much on the score as far as "obsolete architectures" go...witness the incredible survival of the x86. Fundamentally, x86 has a lot of problems because it is building off of a model that was designed over 20 years ago. All that backwards compatibility has really hindered innovation, but nevertheless we see amazing improvements in processor speed every 6 months.

The new Merced processors will (hopefully) put an end to x86, but the road appears rocky. There is such a rich base of programs,too, for x86 that many people will be loathe to switch to something new.

My point is that if Palm were to completely abandon its installed base, this would probably be a bad move. A lot of the benefit of the Palm platform is the software available for it. Emulators are very tricky things, and generally take a lot of CPU power. For Palm to be able to create a new end product that can emulate the Dragonball/EZ and PalmOS 3/4, it would have to be amazingly faster than today's Palms. And still fit the same form factor. And get good batterytime, despite higher clock.

A much better idea,at least in the short term, is probably to keep making incremental OS upgrades, and possibly use a higher clock logic-compatible chip, or a new chip that supports as a subset all of the current assembly calls.

Whew!

Registered: Sep 1999
Location: Milwaukee, WI
Posts: 74

Well I am going to have to post some more. (pretty transparent post hunh )

Registered: Sep 1999
Location: VA, USA
Posts: 27

They could use an "almost DragonBall" CPU. For example the ColdFire has all the DB's instructions except the bit field ones, and I think the bit fields instructions are not commonly used (the C compiler won't gennerate them for example).

The ColdFire runs at up to 200Mhz (and one or fewer cycles per insn rather then 3 or four per instruction).

Failing that one could use one of the many other 68000 CPU's (the DB is a 68010 plus built in pererphrials as far as I can tell). Unfortunitly the 68030 adds some instructions that the ColdFire doesn't have, and the 68060 is slower then the ColdFire, so going this way cuts off ColdFire path.

There are other things they could do, but I think the ColdFire is probbably the best path.

Of corse there is the problem of the Springboard, which is the DragonBall bus plus a few pins. Nothing else is going to use exactly that bus, so they will have to have a whole controler to run the SpringBoard bus. Which will suck a bit. Fortunitly the DragonBall buss isn't all that complex, so it won't be a very complex controler.